Leave one out and k-fold cross validation| using R| cv.glm | train and test data | prediction error

In this video, I show you in R:
1. how to split data into training and test data
2. how to use training data to build the model and test data to check the model
3. how to do cross-validation using cv.glm()
4. how to do k-fold cross validation using cv.glm()
To learn about the idea behind cross-validation, please check out this video: • What is cross validati... It takes you 8 minutes. and it explains the basics of training and testing data, training and testing errors.
1. Why do we need cross-validations?
2. Two methods of cross-validations:
a. leave one out
b. k-fold.
Again, the full R code from this video is available here: github.com/yz-...

[@hecreatescoding]

This was very helpful, thank you!!

[@alessandromorra7797]

It helped me a lot, thank you!

[@jennyshi1330]

It's a wonderful video, which do me a big favor in time! Thanks!

[@datasciencewithyan4124]

Glad it helps!

[@jinli1835]

Very impressive!

[@rafaelpugoni]

How you can do the leave one out cross validation for mixed effects models?

[@gueshmebrahtom3443]

thank you very much , but i encounterd a difficuality to conduct leave one-out sensitivity analysis using R for prevalence or signle proportion meta analysis

[@user-tz3hp6mu6p]

Thanks a ton.

[@datasciencewithyan4124]

Most welcome!

[@manuelleitner1996]

Thank you very much for the detailed and helpful video. Is there a way to interpret the delta value of the cv.glm() function without comparing it with other models (e.g. as done in your video)? For example, can I simply run the LOOCV on a model and then interpret the delta value (if it's good/bad)? Since the MSE depends on the metric/unit, you can't really derive much from this unstandardized value by itself, right? Thank you in advance!!!

[@yuchuxie8170]

Very clear explanation! Learned a lot from your video❥(^_-)

[@melon8419]

Your video helps me a lot. Thanks!!

[@jennyshi1330]

But one question that I encountered in R, is it the same script when the outcome is a binomial variable?

[@datasciencewithyan4124]

I didn’t get your question. Do you want to create a 0/1 binary variable? It would be the same.

[@jennyshi1330]

@@datasciencewithyan4124 Thx, I will try🙏

[@jennyshi1330]

Hello, may I ask how to get the sensitivity and specificity using the LOOCV scripts in your video?

[@a1cswiz1611]

You need to make a new video because your video and audio are out of sync, do you check this stuff before you hit upload?

[@datasciencewithyan4124]

It is not out of sync from my side.